Abstract: Tuberculosis (TB) is an infectious disease of global importance;in 2006, there were more than 9 million incident cases and 1.7 million deaths attributable to TB. The emergence of multidrug-resistant (MDR) and extensively drug-resistant (XDR) TB and the convergence of HIV and TB epidemics are threats to effective TB control. Furthermore, evidence exists that previous M. tuberculosis infection confers limited immunity to re-infection, that an individual can simultaneously harbor more than one distinct strain of M. tuberculosis, that distinct lineages of M. tuberculosis differ in their virulence characteristics, and that M. tuberculosis diversifies within a host. Each of these factors contributes to the within-host complexity of M. tuberculosis infection and presents complications for the treatment of individuals and the control of disease in populations. I propose an observational study among individuals starting treatment for TB in Lima, Peru and Pietermaritzburg, South Africa to evaluate the prevalence, risk factors and consequences of complex M. tuberculosis infection. I will 1) estimate the site-specific prevalence of multiple-strain M. tuberculosis infection and clonal heterogeneity among individuals at the time of initial diagnosis;2) determine the host- and strain-related risk factors for multiple-strain infection and clonal heterogeneity;3) evaluate the effect of multiple-strain infection and clonal heterogeneity on early response to standard first-line treatment regimens;and 4) develop mathematical models to examine the individual- and population-level effects of multiple-strain infection and clonal heterogeneity. Public Health Relevance: Recent studies indicate that M. tuberculosis infections are more complex than had previously been appreciated: an individual can be infected by more than one strain and each strain can evolve during the course of infection. Little is currently known about the prevalence of complex infections, the pathogenand host-factors related to complex infections, or the effect of complex infections on the treatment outcomes of individuals or on the performance of strategies to control disease spread in communities. The studies proposed here will assess the pre-treatment prevalence of complex infections and the impact of complex infection on early treatment response in TB patients in Lima, Peru and Pietermaritzburg, South Africa. These results will be linked to the development of mathematical models designed to assess the effect of complex M. tuberculosis infection on the projected performance of new strategies for TB control. The overall goal of this study is to learn more about the natural history of TB and to use this knowledge to improve the health of individuals and their communities.